Calculate retention by cohort over multiple periods. Track how many customers from each acquisition month remain active in subsequent periods.
Cohort analysis groups customers by their acquisition period (typically the month of first purchase) and tracks what percentage remain active in each subsequent period. It is the gold standard for measuring retention trends because it controls for growth — unlike blended retention metrics that mix new and old customers.
A cohort retention table is read left to right: Period 0 is always 100% (the cohort itself), Period 1 shows what percentage came back the next period, and so on. Healthy businesses show a flattening curve — the customers who survive the first few periods tend to stick around for a long time.
This calculator lets you enter cohort size and active users for up to 6 periods, computing retention percentages and visualizing the decay curve. Use it to compare cohorts, evaluate whether retention is improving over time, and model the long-term value of different customer groups. Whether you are a beginner or experienced professional, this free online tool provides instant, reliable results without manual computation.
Blended retention metrics can look stable while individual cohorts are deteriorating (if you keep acquiring more customers). Cohort analysis reveals the true retention story and lets you evaluate specific changes (new onboarding sequence, loyalty program launch) by comparing before and after cohorts. Having a precise figure at your fingertips empowers better planning and more confident decisions.
Cohort Retention (Period N) = Active Customers in Period N / Cohort Size × 100 Drop-off Rate (Period N) = (Active N-1 − Active N) / Active N-1 × 100
Result: P1: 35%, P2: 20%, P3: 16%
A cohort of 1,000 customers had 350 active in Period 1 (35% retention), 200 in Period 2 (20%), and 160 in Period 3 (16%). The curve is flattening — drop-off from P2 to P3 is only 4 points, suggesting the remaining customers are becoming loyal.
Warren Buffett once said the most important thing is to see what's happening underneath the surface. Cohort analysis does exactly that for customer retention. A business growing at 30% per year might have declining cohort retention that will eventually catch up with it. The cohort table reveals this truth.
The first-month drop is always the steepest because many first-time buyers are one-time experimenters. By month 3–4, you see your "core" retention rate — the percentage of buyers who become genuine repeat customers. This core rate is your true retention baseline.
Launching a new onboarding email sequence? Compare the retention curves of the pre-launch cohort vs. the post-launch cohort at the same maturity. This is the cleanest way to measure the impact of retention programs without confounding factors.
A cohort is typically all customers who made their first purchase in the same month. You can also create cohorts based on acquisition channel, product category, or campaign to compare specific groups.
Period 0 is the cohort itself at 100%. Period 1 is the next time window (e.g., the next month). The retention percentage in each period shows what fraction of the original cohort was still active.
A healthy curve drops steeply initially (many first-time buyers do not return) then flattens. If 15–25% of a cohort remains active after 6 months with minimal further drop-off, those are loyal customers who will likely stay for years.
Track at least 6–12 months for non-subscription e-commerce. The value is in seeing whether the curve flattens. If it does not flatten and keeps declining, you have a fundamental retention/product problem.
Overall retention blends all customers together, masking cohort-specific trends. Cohort analysis isolates each group, letting you see whether newer cohorts retain better or worse than older ones. It is far more actionable for evaluating changes.
GA4 has a built-in cohort exploration report. It is useful for quick analysis. For deeper cohort analysis with revenue and margin data, most teams export data to a spreadsheet or BI tool for custom analysis.